Operators¶
- class rxsci_river.DriftStatus(value)¶
An enumeration.
- class rxsci_river.Prediction(utterance, prediction)¶
The prediction of an utterance
- prediction¶
The actual prediction value
- utterance¶
The utterance associated to the prediction
- class rxsci_river.Utterance(data, label=None, target_values=None)¶
An Utterance value
- data¶
The actual data of the utterance
- label¶
[Optional] The label of the utterance
- target_values¶
[Optional] The list of possible target values
- rxsci_river.compute_metric(metric)¶
Computes a metric on predicted values
- Source:
An observable emitting Prediction items.
- Parameters
metric – A river metric
- Returns
An Observable emitting the metric results for each input prediction.
- rxsci_river.detect_concept_drift(model)¶
Concept drift detection.
- Source:
An Observable of integer or real values.
- Parameters
method – A river drift detection object.
- Returns
An observable emitting DrifStatus items.
- rxsci_river.evaluate.prequential(model, pretrain_size=200)¶
Prequential predict/train evaluation of a model
The source items may or may not have associated labels. If a label is present, then the item is used for training. Otherwise only inference is done.
- Source:
An observable emitting tuples of (x, y) items.
- Parameters
model – A river model object
pretrain_size – [Optional] number of initial items used to train the
before doing predictions. (model) –
- Returns
An Observable emitting prediction items for each input item. The firsts pretrain_size items do not emit predictions.